Adversarial Perturbation Defense on Deep Neural Networks

Deep neural networks (DNNs) have been verified to be easily attacked by well-designed adversarial perturbations. Image objects with small perturbations that are imperceptible to human eyes can induce DNN-based image class classifiers towards making erroneous predictions with high probability. Adversarial perturbations can also fool real-world machine learning systems and transfer between different architectures and datasets. Recently, defense methods against adversarial perturbations have become a hot topic and attracted much attention. A large number of works have been put forward to defend against adversarial perturbations, enhancing DNN robustness against potential attacks, or interpreting the origin of adversarial perturbations. In this article, we provide a comprehensive survey on classical and state-of-the-art defense methods by illuminating their main concepts, in-depth algorithms, and fundamental hypotheses regarding the origin of adversarial perturbations. In addition, we further discuss potential directions of this domain for future researchers.

Who meets national early childhood sleep guidelines in Aotearoa New Zealand? A cross-sectional and longitudinal analysis

Study Objectives To investigate the proportion of children in Aotearoa New Zealand (NZ) who do or do not meet sleep duration and sleep quality guidelines at 24 and 45 months of age and associated sociodemographic factors. Methods Participants were children (n=6,490) from the Growing Up in New Zealand longitudinal study of child development with sleep data available at 24 and/or 45 months of age (48.2% girls, 51.8% boys; 22.4% Māori [the Indigenous people of NZ], 12.9% Pacific, 13.4% Asian, 45.2% European/Other). Relationships between sociodemographic factors and maternally-reported child sleep duration (across 24 hours) and night wakings were investigated cross-sectionally and longitudinally. Estimates of children in NZ meeting sleep guidelines were calculated using a range of analytical techniques including Bayesian linear regression, negative binomial multiple regression, and growth curve models. Results In NZ, 29.8% and 19.5% of children were estimated to have a high probability of not meeting sleep duration guidelines and 15.4% and 8.3% were estimated to have a high probability of not meeting night waking guidelines at 24 and 45 months respectively, after controlling for multiple sociodemographic variables. Factors associated cross-sectionally with children’s sleep included ethnicity, socioeconomic deprivation, material standard of living, rurality and heavy traffic, and longitudinal sleep trajectories differed by gender, ethnicity and socioeconomic deprivation. Conclusions A considerable proportion of young children in NZ have a high probability of not meeting sleep guidelines but this declines across the ages of 24 and 45 months. Sleep health inequities exist as early as 24 months of age in NZ.

Computing top-k temporal closeness in temporal networks

The closeness centrality of a vertex in a classical static graph is the reciprocal of the sum of the distances to all other vertices. However, networks are often dynamic and change over time. Temporal distances take these dynamics into account. In this work, we consider the harmonic temporal closeness with respect to the shortest duration distance. We introduce an efficient algorithm for computing the exact top-k temporal closeness values and the corresponding vertices. The algorithm can be generalized to the task of computing all closeness values. Furthermore, we derive heuristic modifications that perform well on real-world data sets and drastically reduce the running times. For the case that edge traversal takes an equal amount of time for all edges, we lift two approximation algorithms to the temporal domain. The algorithms approximate the transitive closure of a temporal graph (which is an essential ingredient for the top-k algorithm) and the temporal closeness for all vertices, respectively, with high probability. We experimentally evaluate all our new approaches on real-world data sets and show that they lead to drastically reduced running times while keeping high quality in many cases. Moreover, we demonstrate that the top-k temporal and static closeness vertex sets differ quite largely in the considered temporal networks.

Weakly measured while loops: peeking at quantum states

A while loop tests a termination condition on every iteration. On a quantum computer, such measurements perturb the evolution of the algorithm. We define a while loop primitive using weak measurements, offering a trade-off between the perturbation caused and the amount of information gained per iteration. This trade-off is adjusted with a parameter set by the programmer. We provide sufficient conditions that let us determine, with arbitrarily high probability, a worst-case estimate of the number of iterations the loop will run for. As an example, we solve Grover's search problem using a while loop and prove the quadratic quantum speed-up is maintained.

Conjugate and non-conjugate controls of sensitizer to enhance dye-sensitized upconversion luminescence

Although dye sensitization is an efficient strategy to improve upconversion luminescence of lanthanide-doped upconversion nanoparticles (UCNPs), the sensitization effect is limited due to the high probability of dyes aggregation. To...

Improve Neural Distinguishers of SIMON and SPECK

Deep learning has played an important role in many fields, which shows significant potential for cryptanalysis. Although these existing works opened a new direction of machine learning aided cryptanalysis, there is still a research gap that researchers are eager to fill. How to further improve neural distinguishers? In this paper, we propose a new algorithm and model to improve neural distinguishers in terms of accuracy and the number of rounds. First, we design an algorithm based on SAT to improve neural distinguishers. With the help of SAT/SMT solver, we obtain new effective neural distinguishers of SIMON using the input differences of high-probability differential characteristics. Second, we propose a new neural distinguisher model using multiple output differences. Inspired by the existing works and data augmentation in deep learning, we use the output differences to exploit more derived features and train neural distinguishers, by splicing output differences into a matrix as a sample. Based on the new model, we construct neural distinguishers of SIMON and SPECK with round and accuracy promotion. Utilizing our neural distinguishers, we can distinguish reduced-round SIMON or SPECK from pseudorandom permutation better.

Metoclopramide-induced Acute Dystonia Misdiagnosed as an Epileptic Seizure in a Lupus Patient

Acute dystonic reactions are the most prevalent extrapyramidal adverse effects associated with metoclopramide. It could be mistaken for a variety of other conditions, such as seizures, tetanus, and encephalitis, to name a few possibilities. We present a case of a 26-year-old female misdiagnosed as having an epileptic seizure who was rushed to the emergency unit with an involuntary bilateral upward deviation of the eyes, spasm, stiffness, lateral deviation of the neck, and protrusion of the tongue. Symptoms occurred 36 hours after the commencement of metoclopramide, used to treat nausea and vomiting in the referring hospital. All the laboratory work was normal. The drug was discontinued and 5 mg of intravenous biperiden was administered. The symptoms subsided in about 10 minutes with no recurrence. Metoclopramide-induced acute dystonia not only creates an anxious environment for patients but may also be life-threatening. Due to the high probability of misdiagnosis, detailed drug history and a high index of suspicion are critical in making the correct diagnosis.

Changing Regional Inequality Patterns in Western China: A Case Study of Xinjiang

Regional inequality in western China has been rarely examined and has risen as an important issue in the Chinese academic circle. This study, taking Xinjiang as a case, investigated regional inequality in Xinjiang since the 1990s. The results show that the interregional inequality became more prominent, while the overall regional disparities among counties and cities have experienced an inverted U-shape path and have been mainly caused by the inequality of the intra-north part of the province. Counties with high economic development levels were mainly located in northern Xinjiang and had a high probability of further moving economic levels. We further found that the regional development pattern in Xinjiang was significantly associated with spatial accessibility, followed by population density and urbanization factors. This paper deepens our understanding of the complex pattern and mechanism of regional inequality in western China, and local specific policies, therefore, are needed to solve the problem.

Potential pollinators of Fragaria ananassa (Rosales: Rosaceae) in the Perico valleys of Jujuy (Argentina)

The aims of this study were to identify and estimate the potential pollinators of Fragaria ananassa Duch. “strawberry”, through the analysis of pollen carried out on their bodies. The study was carried out on a commercial farm in Las Pampitas, located in the Perico valleys, Jujuy province (Argentina). Floral visitors and other arthropods were sampled and identified, and their diversity was analyzed. The samples of insects selected to obtain the pollen carried were processed following the acetolysis technique. The potential pollinators of the strawberry crop were estimated by calculating the pollination probability index. The floral visitors corresponded to representatives of Hymenoptera, Diptera, Hemiptera, Lepidoptera, Coleoptera and even Araneae. Apis mellifera L. was the most frequent floral visitor and presented more strawberry pollen than other selected insects. The pollination probability index estimated that this species had a high probability of being a pollinator. Other pollinators such as Allograpta exotica Wiedemann, Halictidae, Meliponini and even Thripidae could complement the action of the honeybee. This is the first contribution to the knowledge of the pollination of F. ananassa in the agroecosystems of Jujuy.

CURATIVE BENEFIT OF SUBMUCOPERICHONDRIAL RESECTION OPERATION ON PULMONARY AR

Objective: We aimed at establishing a correlation between deviated nasal septum and pulmonary artery hypertension in patients reporting in Otorhinolaryngology (ENT) clinic at Combined Military Hospital Quetta, and evaluating the beneficial impact of sub mucoperichondrial resection (SMR) operation on right heart myocardial function, primarily related to Pulmonary Arterial Pressures of these patients of longstanding nasal septal deviation. Study Design: Descriptive; hospital-based cross-sectional survey. Place and Duration of Study: Departments of Otorhinolaryngology & Cardiology, Combined Military Hospital, Quetta, from Sep 2019 to Apr 2021. Methodology: Electrocardiographic and 2-Dimensional Echocardiographic parameters of 87 randomized patients suffering from symptomatic longstanding deviated nasal septum (DNS) who had consented to undergo sub mucoperichondrial resection (SMR) operation were compared and studied for any change in status of probability of pulmonary arterial hypertension before and two months after their surgery. Probability of pulmonary hypertension was estimated using probability criteria from updated European Society of Cardiology Pulmonary Hypertension Guidelines 2019. Patients were classified into low, intermediate and high probability depending upon the number of criteria fulfilled by echocardiographic parameters. Results: Two (2.3%) patients suffering from deviated nasal septum presented with p-pulmonale. Twelve (13.8%) reported with right bundle branch block, and 7 (8%) patients demonstrated right axis deviation on electrocardiography. Overall high probability to develop pulmonary hypertension was discovered in 2 (2.3%) patients. We observed a significant improvement in pulmonary artery pressures in patients suffering from long term upper airway obstruction, two months following sub mucoperichondrial resection operation, in terms of maximum velocity and peak tricuspid regurgitation, right ventricle/left..........